ITcon Vol. 20, pg. 193-212, http://www.itcon.org/2015/14

Mapping BIM schema and 3D GIS schema semi-automatically utilizing linguistic and text mining techniques

submitted:July 2014
revised:January 2015
published:January 2015
editor(s):Turk Z
authors:Jack C.P. Cheng, Assistant Professor
The Hong Kong University of Science and Technology
cejcheng@ust.hk

Yichuan Deng, PhD Candidate
The Hong Kong University of Science and Technology
ycdeng@ust.hk

Chimay Anumba, Professor
The Pennsylvania State University
anumba@engr.psu.edu
summary:The interoperability between BIM (Building Information Modeling) and 3D GIS (Geographic Information System) can enhance the functionality of both domains. BIM can serve as an information source for 3D GIS, while 3D GIS could provide neighboring information for BIM to perform view analysis, sustainable design and simulations. Data mapping is critical for seamless information sharing between BIM and GIS models. However, given the complexity of todayÕs BIM schemas and GIS schemas, the manual mapping between them is always time consuming and error prone. This paper presents a semi-automatic framework that we have developed to facilitate schema mapping between BIM schemas and GIS schemas using linguistic and text-mining techniques. Industry Foundation Classes (IFC) in the BIM domain and City Geography Markup Language (CityGML) in the GIS domain were used in this paper. Entity names and definitions from both schemas were used as the knowledge corpus, and text-mining techniques such as Cosine Similarity, Market Basket Model, Jaccard Coefficient, term frequency and inverse document frequency were applied to generate mapping candidates. Instance-based manual mapping between IFC and CityGML were used to evaluate the results from the linguistic-based mapping. The results show that our proposed name-to-definition comparison could achieve a high precision and recall. Results using different similarity measures were also compared and discussed. The framework proposed in this paper could serve as a semi-automatic way for schema mapping of other schemas and domains.
keywords:Building information modeling (BIM), City Geography Markup Language (CityGML), Industry Foundation Classes (IFC), Interoperability, Schema mapping
full text: (PDF file, 0.996 MB)
citation:Cheng JCP, Deng Y, Anumba C (2015). Mapping BIM schema and 3D GIS schema semi-automatically utilizing linguistic and text mining techniques, ITcon Vol. 20, pg. 193-212, https://www.itcon.org/2015/14